Talk:Russell1930 5

From Embryology

V THE THEORY OF THE GENE[edit] W EISMANN held that each chromosome must contain at least one full set of determinants (and in most cases several sets). He arrived at this conclusion in a perfectly logical way. He argued that if the chromosomes were different, so that each held only the determinants for a particular part of the body or for a particular set of characters, it would be difficult or impossible for the fertilized egg to receive invariably, as clearly it does receive, a complete set of determinants. It would be necessary to assume, for instance, that at least the haploid set of chromosomes contained the determinants for the whole organism, for if it did not, but bore only those for certain parts, it would be a rare occurrence that the egg and sperm uniting to form the zygote carried exactly complementary groups of determinants and a perfect embryo resulted. But each haploid group contains chromosomes of maternal and of paternal origin, which we have no justification for supposing to be always complementary to one another; hence it seems to follow that, in order to account for the fact that a complete organism invariably arises from the fertilized egg, each chromosome must contain all the determinants necessary for complete development, and the fertilized egg many such sets of determinants . 1

But Weismann overlooked one possibility — that prior to reduction the homologous maternal and paternal chromosomes should pair and that at reduction they should separate, so that each gamete would receive a complete haploid set of chromosomes, as a rule of mixed paternal and maternal origin, and each zygote two complete sets.

Near the beginning of the present century cytological research showed that this possibility was actually realized in certain cases, and the view that reduction was preceded by conjugation of homologous chromosomes of maternal and

1 V ortrage, i, pp. 378-81 (Eng. Trans., pp. 345-8).


paternal origin was rapidly generalized and widely accepted. About the same time, Boveri’s classical experiments with dispermic sea-urchin eggs furnished clear evidence that the chromosomes were in fact qualitatively different — as indeed Roux (1883) 1 had deduced on theoretical grounds — and that for normal development one complete haploid set was the minimum necessary. The other factor which, combined with the new cytological discoveries, led to the theory of the gene, was of course the re-discovery in 1900 of the Mendelian principles of heredity.

Sutton (1902 and 1903) appears to have been the first to draw the complete parallel between the cytological facts and the Mendelian results, and to explain Mendelian segregation as due to the separation of maternal and paternal chromosomes at reduction and their random distribution to the gametes. It is useful here to quote Wilson’s account of Sutton’s views:

‘In brief summary Sutton’s conclusions were as follows:

‘(i) The somatic or diploid chromosome-groups are made up of two equivalent chromosome-groups or series, one of maternal derivation and one of paternal (Van Beneden, Boveri, Montgomery).

‘(2) The chromosomes retain their morphological individuality and are genetically continuous throughout the life-cycle (Van Beneden, Rabl, Boveri).

‘(3) The process of synapsis consists in the union or conjugation of corresponding or homologous maternal and paternal chromosomes which in the reduction division disjoin, pass to opposite poles of the spindle, and thus always into different germ-cells (Montgomery, Sutton).

(4) Each chromosome plays a definite part in the determination of development (Boveri).’

‘To the foregoing Sutton added the following theoretic postulates:

‘(5) A given size-relation is characteristic of the physical basis of a definite set of genetic units. Each chromosome of any haploid series in the species has a homologue in any other series, and these homologous members of each pair “cover the same field” in development.

1 Ueber die Bedeutung der Kerntbeilungsfiguren , Leipzig, 1883. Reprinted in Gesammelte AbbandL (No. 17), Leipzig, 1895.


This means, in more modern terminology, that the synaptic mates contain the physical units (factors or genes) that correspond to the Mendelian allelomorphs.

(6) “In the reduction-division the position of the chromosomepairs or bivalent chromosomes in the equatorial plate is purely a matter of chance — that is, any chromosome-pair may lie with maternal or paternal chromatid toward either pole irrespective of the positions of other pairs — and hence a large number of different combinations of maternal and paternal chromosomes are possible in the mature gametes of an individual. ,,

‘Sutton clearly showed that the formulas of Mendelian heredity generally (as then known) could be applied without alteration alike to the hypothetical “factors” or “genes” and to the chromosomes; and that the combinations, segregations, and recombinations of the former are paralleled by those of the latter. Sutton also foreshadowed the modern theory of linkage, pointing out the necessity for the assumption that “some chromosomes at least are related to a number of different allelomorphs”; that “all the allelomorphs represented by any one chromosome must be inherited together”; and further, that “the same chromosome may contain allelomorphs that may be dominant or recessive independently” (’03, p. 240). This, as far as it goes, is in all essentials identical with the results afterwards worked out by Morgan and his co-workers in genetics which demonstrate that in the case of Drosophila at least the number of linkage groups is the same as the haploid number of chromosomes .” 1

The extensive and elaborate breeding experiments carried out from about 1910 onwards by Morgan and his school disclosed many interesting facts about the inheritance of a large number of mutations occuring in the fruit-fly Drosophila, and enabled the results to be expressed in numerical ratios. For the interpretation of these results Morgan utilized and extended the conception of Sutton that Mendelian inheritance is bound up with the distribution of the chromosomes in reduction.

It is not necessary for our special purpose here to enter into details regarding Morgan’s theory of the gene, nor to concern ourselves with the ingenious hypotheses of crossing-over, interference, lethal factors, and so on, by means of which

1 Wilson, 1925, pp. 926-7.


deviations from the expected numerical ratios are explained . 1 What we are concerned with are the fundamental postulates of the theory and the justification for them.

Morgan sums up the essentials of the theory as follows : 2

‘The theory states that the characters of the individual are referable to paired elements (genes) in the germinal material that are held together in a definite number of linkage groups; it states that the members of each pair of genes separate when the germ-cells mature, in accordance with Mendel’s first law; it states that the members belonging to different linkage groups assort independently in accordance with MendePs second law; it states that an orderly interchange — crossing-over — also takes place, at times, between the elements in corresponding linkage groups; and it states that the frequency of crossing-over furnishes evidence of the linear order of the elements in each linkage group and of the relative position of the elements with respect to each other.’

Another protagonist of the gene theory is even more sweeping in his statements.

‘The hereditary constitution of at least all higher organisms’, writes Professor J. S. Huxley, 3 ‘consists of a number of units (factors or genes), each of which may exist in a number of forms (allelomorphs); these genes exist in definite proportions, and are arranged in a definite order; the whole gene-complex is divided up among the separate chromosomes, which in Drosophila have been shown to correspond to the linkage-groups established by genetic experiments.’

It would seem at first sight that we have here to do with a mere variant of the Weismannian idea of determinants, that the genes or factors are the sole determiners of the characters whose names they bear, and there is in fact a close analogy between the two concepts. The existence of genes is however deduced on quite different grounds, and Morgan

1 These are critically dealt with by H. Stieve, ‘Ncuzeitliche Ansichten uber die Bedeutung der Chromosomen, unter besonderer Beriicksichtigung der Drosophila versuchc*, Ergebn.Anat. Entwtck. xxiv, 1923, pp. 491-587. See also J. Dembowski, ‘Zur Kritik der Faktoren- und Chromosomenlehre’, Zts. indukt . Abstamm. Vererbungslebre , xli, 1926, pp. 216-47.

3 T. H. Morgan, The Theory of the Gene , New Haven, 1926, p. 25.

3 Nature , Dec. 25th, 1926, p. 903. See also H. S. Jennings, Prometheus , London, n.d., for a bold and picturesque account of the theory.


specifically denies that they can be regarded as equivalent to representative particles.

When a female Drosophila with red eyes is mated with a male having white eyes (a mutation), the F i generation all have red eyes, and the F 2 generation show a ratio of three reds to one white — an ordinary case of Mendelian inheritance. It can be interpreted on the assumption that in the F I hybrid ‘something for red eyes has separated from something for white eyes’. We are told also that ‘We may express these factorial relations in another way by saying that a germ cell that produces white eyes differs from a germ cell that produces red eyes by one factor-difference. We think of this difference as having arisen through a factor in the red-eyed wild fly mutating to a factor for white.’ 1

Now the interpretation offered in this last sentence does not seem to be a necessary deduction from the facts, and it reveals to us the fundamental assumption or postulate of the whole theory. Let us accept the evidence that the mechanism of Mendelian inheritance is bound up with the chromosomes ; let us further agree that the chromosome A , which is associated with the production of red eye, differs from the chromosome A', which is in some way responsible for the appearance of white eye. The facts appear to justify, or at least to be consistent with, these conclusions. What is the nature of the difference between these two chromosomes ? Morgan assumes that the difference is particulate, that one locus of chromosome A' differs from the corresponding locus of chromosome A, and that these loci can in practice be treated as detachable and movable material units. We get at once from this primary assumption the whole elaborate theory of the gene. But the difference between the two chromosomes may quite well be a slight though discontinuous chemical or stereochemical change affecting the chromosome as a whole. 2 We do not ascribe the difference in properties

1 Morgan, Sturtcvant, Muller, and Bridges, T be Mechanism of Mendelian Heredity , revised edition, New York, 1922, p. 262.

2 Equally well it may affect principally one particular section or locus of the chromosome (see below, p. 281). The point is that it should not be hypostatized as a definite and independent particulate unity which is necessarily a pure abstraction.


between two chemical isomers to a f articulate difference in their constitution, but to a different arrangement of their constituent atoms. Nor do we attempt to pin down the difference in properties to one constituent of the molecule . 1 Why should we then adopt the particulate explanation in the case of differing chromosomes I

The original ground for assuming that the difference is a particulate one appears to be that the effect shown by the developed fly is a definite and discontinuous change manifested primarily in one character. It is natural to think that a discontinuous effect of this kind has a particulate cause in the germ-cell. The gene theory in its original form was linked up with the concept of unit characters. The process of thought is the same as that which led Weismann to postulate the existence of separate determinants to account for the independent heritability of small definite, discontinuous characters. But the conception does not emerge inevitably from the facts — another mode of explanation is equally possible. The gene theory is therefore only a hypothesis, and the existence of genes as concrete things cannot be certainly demonstrated from arguments of this character.

It is interesting to note that the development of genetic research has shown that the relation of gene to character is not the simple one of determinant to determinate, and the concept of unit characters is accordingly losing its importance.

‘Mendelian heredity has taught us that the germ cells must contain many factors that affect the same character. Red eye color in Drosophila, for example, must be due to a large number of factors, for as many as 25 mutations for eye color at different loci have already come to light. Each produced a specific effect on eye color; it is more than probable that in the wild fly all or many of the normal allelomorphs at these loci have something to do with red eye color. One can therefore easily imagine that when one of these 25 factors changes, a different end result is produced, such as pink eyes, or vermilion eyes, or white eyes or eosin eyes. Each such color may be the product of 25 factors (probably of many more) and each set of 25 or more differs

1 See Delage (below, p. 82).


from the normal in a different factor. It is this one different factor that we regard as the “unit factor” for this particular effect, but obviously it is only one of the 25 unit factors that are producing the effect. However since it is only this one factor and not all 25 which causes the difference between this particular eye color and the normal, we get simple Mendelian segregation in respect to this difference.’ 1

Conversely, any one factor may affect more than one character; that for rudimentary wings may affect also the legs, the number of eggs laid, and the general viability of the organism.

‘The genetic evidence’, writes Morgan in his latest book, 2 ‘has abundantly shown that when a single gene is changed, the end-product may be affected in many ways. We select the most important, and for our purposes, most convenient change and identify it as the immediate product of the new gene. But as is well understood, this is only a useful method in studying the inheritance of the genes. It is perfectly well known that, besides certain major effects, there are many accompanying effects also present involving all parts of the body, and this result is entirely consistent with the theory that all the parts are the products of all the genes.’

It is clear then that there is no simple relation between character and factor, and the argument from the genetic behaviour of the character to the existence of the factor or gene loses much of its force.

Throughout the whole of the Drosophila experiments, we have to do with the hereditary behaviour of small distinct differences (mutations) from the normal. There is much solid evidence that these differences are associated with differences between pairs of chromosomes.

To quote Morgan again:

‘The germ cells may be thought of as a mixture of many chemical substances, some of them more closely related to the production of a special character, color for example, than are others. If any one of the substances undergoes a change, however slight, the end-product of the activity of the germ cell may be different. All sorts of characters might be affected by the change, but certain parts might be more

1 Morgan and others, op. cit., pp. 262-3.

3 Experimental Embryology , New York, 1927, p. 207.


conspicuously changed than are others. It is these more obvious effects that we seize upon and call unit characters.’ 1

That is a reasonable and objective statement, but the gene theory goes very much farther. In effect it ‘reifies’ or endows with material existence what are merely differences, and it does this by postulating a gene for every heritable difference found. It may be said that this is an unfair statement, and that the gene for each atypical character, each mutation, is merely a modification or mutation of the gene for the typical character, 2 and does not represent the modification per se. But this merely means that a normal gene has already been postulated to account for the typical character — and both genes are equally hypothetical.

The postulation of separate genes for each distinguishable difference and of the equivalent normal genes leads to very great complications. Take the case, described in the quotation on p. 59 above, of the twenty-five factors required for the production of normal eye-colour in Drosophila. Let us grant that there are in fact twenty-five different modifications of the various chromosomes concerned, each of which can bring about a definite, observable, and constant abnormality in eye-colour. The facts may legitimately bear that interpretation (though it is difficult to concede that twentyfive distinct shades of colour can be accurately distinguished in the eye of Drosophila ). Is it really necessary to assume that each of these differences is localized in a particular section of the chromosome in the shape of a mutant gene, and that it has a counterpart in the ‘ normal ’ chromosome in the shape of a normal gene f This is surely a breach of the good old medieval rule that ‘Entia non sunt multiplicanda praeter necessitatem’. The result obtained — that at least twenty-five factors or genes (whose nature is completely unknown, and whose relation to ascertained physiological processes is completely

1 Morgan and others, op. cit., p. 264.

1 Cf. F. A. E. Crew : ‘A gene is a particular state of organization of the chromatin at a particular point along the length of a particular chromosome. It is a particular area or locus of the chromosome in a particular state. One particular condition of this chromatin can be replaced by others and with each change another gene appears', Nature , Nov. 19th, 1927, p. 733.


mysterious) are concerned in the production of normal eyecolour — is frankly incredible. Eye-colour in Drosophila is, physiologically speaking, a resultant of the mixture of two pigments only — wine-red and ochre-yellow. ‘The eyes of the eye-color mutants, “eosin-miniature”, “eosin-vermilion”, “pink”, “purple”, “ruby”, “sepia”, “tinged”, “vermilion”, and “white”, have the same structures and pigments possessed by the normal eye, differing only in the amount and distribution of the pigment. The pigments are much reduced in “tinged”, “eosin-vermilion”, and in the female of “eosin-miniature”.’ 1 It seems highly doubtful that such an elaborate mechanism of mutant genes is necessary to induce such relatively simple changes, and it borders on the miraculous that the production of the normal amount and distribution of the two pigments concerned should require the combined action of twenty-five ‘and probably many more’ normal genes. Generalizing from this example, it would seem that there must exist as many normal genes as there are separately distinguishable characters, and as many mutant genes as there are mutations of these characters, that is to say, an unlimited number of genes. The hypothesis crumbles under its own weight.

We are not, however, primarily concerned here with a detailed criticism of the gene theory, especially as it has clearly not yet reached the term of its evolution. Our object has merely been to show that the genes are purely hypothetical units — convenient (or inconvenient ?) fictions invented to account for the very complex hereditary behaviour of mutant characters in Drosophila. As such they have a certain interpretative and heuristic value, provided that their purely conceptual and hypothetical character is clearly borne in mind. That the Drosophila results are interpretable on a pure and abstract factorial hypothesis, quite apart from any identification of the factors with material and unitary parts of the chromosomes, is of course freely admitted by Morgan, who writes, ‘The factorial theory as such deals with the behavior

1 O. A. Johannsen, ‘Eye Structure in Normal and Eye-mutant Drosophila*, Journ. Morfb . Physiol. y xxxix, 1924, pp. 337-50.


of its factors in an abstract way, quite apart from any material basis of which they may happen to be composed. In this way, it may measure their constancy, segregation, linkage, & c .’ 1

To arrive at a just estimate of the value of the gene theory in relation to the major problems of development and heredity, we cannot do better than follow the lead of the late Professor W. Johannsen, himself the inventor of the term ‘gene’, and a most distinguished worker in genetics. In a short but fundamental paper 2 he lays his finger on the limitations of modern genetical theory, and assesses its general significance in masterly fashion. We shall here reproduce his views in some detail.

After pointing out that the theories of heredity propounded by Darwin, Galton, Weismann, De Vries, and the early Mendelians were profoundly morphological, dealing as they did with parts of the organism regarded as units in inheritance, he shows that the later concept of unit characters is also a morphological notion, implying a reification of qualities, and can now no longer be upheld. What is required is a physiological or ‘chemico-biological’ formulation of the facts.

‘It was undoubtedly a step forward to leave the notion of \mit-parts in favour of the notion of unit -characters. Now this notion too is absolutely untenable. Nowadays each of Bateson’s allelomorphs are not regarded as a kind of germ (“Anlage”) for a corresponding unitcharacter. My term “gene” was introduced and generally accepted as a short and unprejudiced word for unit-factors in the — as to heredity — essential constitution of gametes and zygotes, but originally I was somewhat possessed with the antiquated morphological spirit in Galton’s, Weismann’s and Mendel’s viewpoints. From a physiological or chemico-biological standpoint we must a priori in characters or developed parts of organisms see Reactions of the (I should say genotypical) constitution belonging to the zygote in question; and from this point of view there are no unit-characters at all ! Undoubtedly all scientific geneticists now are or ought to be in accord as to this matter* (p. 136).

In genetics we have to do

‘with such genotypical units as are separable, be it independently

1 Morgan and others, op. cit., p. 278.

1 ‘Some Remarks about Units in Heredity*, Hereditas , iv, Lund, 1923, pp. 133-41.


or in a more or less mutual linkage. Certainly by far the most comprehensive and most decisive part of the whole genotype does not seem to be able to segregate in units ; and as yet we are mostly operating with “characters” which are rather superficial in comparison with the fundamental Specific or Generic nature of the organism. . . . We are very far from the ideal of enthusiastic Mendelians, viz. the possibility of dissolving genotypes into relatively small units, be they called genes, allelomorphs, factors or something else. Personally I believe in a great central “something” as yet not divisible into separate factors. The pomace-flies in Morgan’s splendid experiments continue to be pomaceflies even if they lose all “good” genes necessary for a normal fly-life, or if they be possessed with all the “bad” genes, detrimental to the welfare of this little friend of the geneticists.

‘Disregarding this (perhaps only provisional ?) central “something” we should consider the numerous genes, which have been segregated, combined or linked in our modern genetic work. What have we really seen ? The answer is easily given : We have only seen Differences . The famous relation 311(1:2:1) indicates one single point of difference, the ratio 9:3:311 two points, and so on. Dominance does not at all indicate the presence of some positive unit, just as little as Recessivity indicates the lack of any unit. This is clearly seen, for instance, in Nilsson-Ehle’s oats-crossings, where one Mendelian unit may be responsible for one dominant and one or two recessive characters, also in such cases where dominance or recessivity is dependent upon external conditions, as in some Drosophila-experiments’ (p. 137).

Further, ‘When we regard Mendelian “pairs”, Aa, Bb and so on, it is in most cases a normal reaction (character) that is the “allel” to an abnormal ’ (p. 138).

As to the nature of the difference between the pairs:

‘There is at present scarcely any doubt about the theory, that “Mendelian” factors are in some way bound in or to the chromosomes. The morphological view regards them as formed particles (say “morphs” ad modum “allellomorphs”) of the chromosomes, an old Weismannian idea — mutatis mutandis. From a physiological standpoint we may prefer to regard local conditions (say “chemisms”) in or on the chromosomes as responsible for those units’ (p. 138).

Finally, Johannsen raises a serious doubt as to the evolutionary significance of the Mendelian phenomena :

‘To my mind the main question in regard to these units is this: Are the experimentally demonstrated units anything more than ex


pressions for local deviations from the original (“normal”) constitutional state in the chromosome ? Is the whole of Mendelism perhaps nothing but an establishment of very many chromosomical irregularities, disturbances or diseases of enormously practical and theoretical importance but without deeper value for an understanding of the “normal” constitution of natural biotypes ? The Problem of Species, Evolution, does not seem to be approached seriously through Mendelism nor through the related modern experiences in mutations’ (p. 140) .

There are two points deserving of special notice in this considered judgement of Johannsen’s: (1) that Mendelian heredity has to do with the inheritance of differences from the normal, mostly superficial differences and often degenerative or pathological, and (2) that in addition to these differential characters, which mendelize, there is the whole body of the main characters of the organism, forming a central group which is apparently transmitted en bloc. Let us consider these points in further detail.

(1) That the Mendelian characters are mainly superficial, affecting the coloration of the organism, or relating mainly to slight bodily peculiarities, is to some extent admitted by Morgan and his school, but they in no way agree that the scope of the Mendelian principle is limited to superficial characters.

‘It is true’, writes Morgan, ‘that by far the greatest number of characteristics that students of Mendelian inheritance have concerned themselves with relate to superficial differences, such as shades of color or slight differences in the length or breadth of characters that are relatively unimportant for the individual. The explanation of this procedure is obvious enough, for it is just these slight differences that do not interfere with the survival of those individuals that are necessary to the geneticists. But there is no line that can be drawn between these trivial differences and those that are more significant and fundamental. Even such a fundamental property as symmetry has been shown to depend on a single Mendelian gene, as when the bilateral flower of the snapdragon changes to a peloric flower with radial symmetry. So far as Mendelian factors are concerned, the evidence is quite sufficient to show the erroneousness of the view that Mendelian genes are concerned only with trifling differences’ (p. 727). 1

1 ‘Mendelian Heredity in relation to Cytology*, in Cowdry, General Cytology , Chicago, 1924.


The point is discussed at some length in his book 7 he Physical Basis of Heredity (Philadelphia and London, 1919), and it is worth while to quote the relevant passage in full :

‘It has been sometimes stated’, he writes, ‘usually by the opponents of Mendel’s theory, or by advocates of doctrines of evolution that appeared to be compromised by the Mendelian conception of “unit factors”, that Mendelism deals only with such superficial characters as the color of flowers or the hair color of mammals. This statement contains an element of truth in so far as it covers most of the kinds of characters that students of heredity find most convenient to study; but it contains an entirely false inference as to the limitations of Mendelism. The issue involved is this : changes in superficial characters are not so likely to affect the ability of the organism to survive as are changes in essential organs ; hence they are the best kind of hereditary characters for study. But there is no evidence that such superficial characters are inherited in a different way from “fundamental” characters, and there is evidence to the contrary. A common class of characters showing perfect Mendelian behavior are so-called lethals that destroy the individual when in homozygous condition. There can be no question as to the fundamental importance of such factors. Between these extreme cases and the superficial shades of eye color, for example, all possible gradations of structure, physiological and pathological, are known. The only possible question that might be seriously raised is whether these characters are all losses or deficiencies, while progressive advances may belong to a different category. This may be a serious question for the evolutionist, but has nothing to do with the problem that concerns us here’ (p. 36).

It is admitted then, that with the exception of the lethal factors, the characters showing Mendelian behaviour are minor characters. The reason for this is, however, alleged to be that these are the most convenient characters to study, since mutations of greater scope are not as a rule viable.

But there is a much more obvious reason for the limitation of the Mendelian principle. Mendelian heredity is shown only in sexual reproduction, and it can be demonstrated only if there are one or more differences between the conjugating gametes: obviously it cannot occur at all if fertile union between the divergent gametes is impossible. The experi


mental demonstration of the principle is therefore limited to cases where the difference between the parental types is not too great to prevent successful fertilization and normal development. It relates only to hereditary differences , and these differences can by the nature of the case be only differences so unimportant that they do not interfere with normal crossing. I say ‘unimportant’ rather than ‘slight’, for quite striking differences can be inherited in Mendelian fashion, e.g. dwarfness or tallness in peas; such differences are, however, never so considerable as to be developmentally incompatible with the realization of the specific form. The argument from lethal factors advanced by Morgan is not in the least convincing. The lethal factor is a figment invented to account for the non-appearance of certain factorial groupings, and the occurrence of such abortive broods or parts of broods (in cases where they are demonstrated to occur) merely illustrates the fact that certain combinations are infertile or non-viable, presumably because they include factors mutually incompatible. Dembowski’s comments on the subject of lethal genes are very much to the point:

‘The lethal factors especially’, he writes, ‘help us to recognize the true nature of the gene. Death is an extremely complicated phenomenon, which may depend upon an innumerable host of circumstances of the most diverse kind. When the cause of death is simply called a “lethal gene”, this shows us clearly what a gene really is. We have already come to the conclusion that a gene cannot be material. Now we have the solution that fulfils that condition. The gene is a word , which enables a complicated happening to be briefly denominated’ (op. cit., p. 244).

We may safely say then that the Mendelian principle, so far as the definite experimental evidence goes, is necessarily limited to the narrow field of fertile inter-crossing. Such differences as are not incompatible with fertile crosses may with justice be called ‘unimportant’ or ‘superficial’ in relation to the normal specific form and characteristics.

It is interesting to note that Morgan is inclined to admit that many of the differences whose hereditary behaviour has been studied on Mendelian lines, especially in Drosophila, are


deficiency variations, 1 and after studying the pictures of the semi-aborted and abnormal flies which make up the bulk of his mutations, one can only agree with him. In general, as he himself says, the mutant types of Drosophila are weaker and less well adapted than the normal wild type. 2

The extension of the Mendelian principle to characters other than those which can be studied in actual crosses is of course based on the general conception that the chromosomes constitute the physical basis of heredity. The argument runs somewhat as follows. The Mendelian factors are certainly borne by the chromosomes, and their behaviour can be interpreted by what we know of the distribution of the chromosomes in maturation, reduction, and fertilization. The chromosomes appear to be the only possible vehicles for the transmission of the characters of the organism as a whole, since they are the only equivalent structures in the male and female gametes, and the contribution of both parents to the hereditary equipment of the offspring appears to be equal. Hence it would seem that what is true of one set of characters should be true of the others, since all are borne by the chromosomes. We arrive therefore at the conception of a genetic constitution, consisting of a large number of factors, like those demonstrated by Mendelian research, and responsible between them — with the co-operation of the cytoplasm and of environmental conditions — for the inheritance and development of all the characters of the organism.

There is, however, contained in this argument a curious petitio principii, which has as a rule escaped notice, though recently it has been pointed out by Winkler. 3 The logical slip is the assumption that the paternal and maternal contributions are of equal importance. For how can this be proved ? In all breeding experiments we can deal only with differences, and only with such differences as are not incompatible with fertile inter-crossing. The great bulk of characters is necessarily common to both parents, and about

1 See quotation on p. 66. 2 Morgan, 1926, p. 65.

3 H. Winkler, *Ueber die Rolle von Kern und Protoplasma bei der Vererbung',

Zts. Indukt . Abstamm. Vererbungslebre , xxxiii, 1924, pp. 238-52.


each parent’s contribution to the inheritance of what is common to both, breeding experiments can obviously tell us nothing. Experimental work can deal only with the inheritance of minor characters or of minor variations of major characters. Accordingly, so far as the facts go and strict logic carries us, inheritance of the main characters might be purely maternal, and involve not only the nuclear apparatus but very intimately the cytoplasm, which is practically absent in the majority of male gametes. The facts under consideration do not of course prove this unilateral inheritance, but they can certainly not be used to demonstrate that the main characters are derived in approximately equal measure from both parents . 1 This being so, the logical argument for the extension of the gene theory to all characters of the organism falls to the ground, and the question remains completely open.

(2) Here then is the point of Johannsen’s suggestion, that in addition to all the separable, mendelizing characters, which can be treated in terms of separate factors, there must be postulated a ‘great central “something”, as yet not divisible into separate factors’. That it will ever be so divisible seems highly improbable.

We must conclude then, until further evidence is forthcoming — if it ever is — that the scope both of the Mendelian principle and of the gene theory is limited to such rather superficial variations as can be dealt with by experimental breeding. The direct contribution which genetic experiment and genetic theory make to the main problem of heredity — the reproduction of specific type, apart from minor deviations — is therefore a strictly limited one; the fundamental problem is in fact hardly touched.

To the problems of development, of individual ontogeny, the gene theory admittedly makes no contribution at all; rather does it add complications. Let us hear Morgan on the subject:

‘Between the characters, that furnish the data for the theory, and the postulated genes, to which the characters are referred, lies the 1 This important point is more fully developed in Chapter XIV below.


whole field of embryonic development. The theory of the gene, as here formulated, states nothing with respect to the way in which the genes are connected with the end-product or character.’ 1

Or again,

‘The cause of the differentiation of the cells of the embryo is not explained on the factorial hypothesis of heredity. On the factorial hypothesis the factors are conceived as chemical materials in the egg, which, like all chemical bodies, have definite composition. The characters of the organism are far removed, in all likelihood, from these materials. Between the two lies the whole world of embryonic development in which many and varied reactions take place before the end-result, the character, emerges. Obviously, however, if every cell in the body of one individual has one complex, and every cell in the body of another individual has another complex that differs from the former by one difference, we can treat the two systems as two complexes quite irrespective of what development does, so long as development is orderly.’ 2

The gene theory deals with the explanation of differences between two or more ontogenies, not with development itself. The gene theory is therefore in no way a theory of development. The introduction of the gene concept into the general theory of development is likely to lead merely to confusion, since, as we have seen reason to conclude, the gene is a purely hypothetical unit, having like Weismann’s determinants no corporeal existence, and invented for a very different purpose than the explanation of the developmental process.

It is in fact peculiarly difficult to harmonize the modern theory of the germ-plasm with the facts of embryonic differentiation, for according to that theory the germ-plasm is the whole chromosomal complex, and this complex occurs apparently unchanged in every cell of the body. Weismann’s theory of a qualitative division of the nucleus and an orderly disintegration of the formative germ-plasm was much more appropriate, but unfortunately it was definitely negatived by the facts. There seems no escape from the conclusion that the chromosomal apparatus is divided only quantitatively, and is handed on unaltered at every mitosis. The chromo 1 Morgan, 19 26, p. a 6. * Morgan and others, 1922, p. 280.


somes appear therefore frima facie to have nothing whatever to do with cellular differentiation, which in fact affects solely the cytoplasm.

Attention has been called to this difficulty particularly by Dobell, with special reference to the cycle of changes exhibited by the life-history of the sporozoon Aggregata eberthi J He writes:

‘In Aggregata we see great individual diversity associated with apparently complete identity of chromosomic constitution. From generation to generation the forms and functions of the animal change in an orderly sequence, while the chromosomes remain unchanged. They are the constants in a varied series of developmental stages. It is therefore obvious that if any internal “factor” in such a sequence of forms “determines” the manifestation of any particular bodily character at any stage, this factor must be somehow associated not with the chromosomes but with some extra-chromosomic constituent of the organism. The organism models itself and acts not because of its chromosomic components but in spite of them.

‘As there is thus no reason whatsoever to suppose that any chromosome — or part of a chromosome — in Aggregata is correlated with or determines the manifestation of any character in the individual actually possessing it, it follows that there is no justification for the further supposition that the chromosomes are specially concerned in the hereditary transmission of any character from any individual to its progeny’ (p. 184).

The same arguments apply also to differentiation in the tissue cells of Metazoa.

The relation of the modern germ-plasm theory to the physiology of development is also discussed — and in a very able way — by F. R. Lillie , 2 who writes :

‘It is apparently not only sound, but apparently almost universally accepted genetic doctrine to-day that each cell receives the entire complex of genes. It would, therefore, appear to be self-contradictory to attempt to explain embryonic segregation by behavior of the genes which are ex hyp . the same in every cell’ (p. 365).

There is a complete divergence between the methods and

1 Clifford Dobell, ‘The Chromosome Cycle of the Sporozoa considered in Relation to the Chromosome Theory of Heredity’, La Cellule, xxxv, 1925, pp. 167-92.

‘The Gene and the Ontogenetic Process’, Science , lxvi, 1927, pp. 361-8.


postulates of genetics and the physiology of development, which is not likely to result in the victory of the geneticists.

‘The present postulate of genetics is that the genes are always the same in a given individual, in whatever place, at whatever time, within the life-history of the individual, except for the occurrence of mutations or abnormal disjunctions, to which the same principles then apply. The essential problem of development is precisely that differentiation in relation to space and time within the life-history of the individual which genetics appears implicitly to ignore. The progress of genetics and of physiology of development can only result in a sharper definition of the two fields, and any expectation of their reunion (in a Weismannian sense) is in my opinion doomed to disappointment. Those who desire to make genetics the basis of physiology of development will have to explain how an unchanging complex can direct the course of an ordered developmental stream* (P- 367) A special study has been made by Hance 1 to determine whether any definite relation can be observed between differentiation of cell and differentiation of chromosomal complex — with definitely negative results.

‘The morphological data’, he writes, ‘on the behavior of the chromosomes in the developing embryo, although admittedly scant, has [sic\ so far given no clue to the manner in which the chromosomes may contribute their potentialities to the growing organism. Before a study of somatic chromosomes had been made it seemed reasonable to expect to find the various highly differentiated cells of the body with chromosome numbers, morphology or behavior at variance both with those found in other tissues and with the specific number and general characteristics found in the gonads. This has been found not to be the case in at least three forms, the pig, the evening primrose and the chick. ... In general the chromosome situation in the soma seems to be entirely similar to that found in the unreduced gonad cells* (p.445).

This is very difficult to explain on the theory that the chromosomes are the structures mainly responsible for heredity and development, and Hance*s tentative hypothesis quoted below sounds very much like an attempt to bolster it up at all costs.

‘In view of the entire similarity*, he writes, ‘of the somatic and

1 R. T. Hance, The Chromosomes of the Chick Soma’, Biol, j Bull., li, 1926, pp. 443-8.


germinal mitotic behavior and in consideration of the complete inability of highly specialized cells to regenerate other than cells similar to themselves, it is tentatively suggested as a basis for future discussion that the somatic chromosomes, as far at least as their genetic function is concerned, have either become functionless or their cytoplasmic environment is incapable of reacting to the possibilities presumably carried by them’ (p. 446).

It would be more logical to conclude that the cause of cellular differentiation lay in the cytoplasm or more properly in the complete cell, in its relations to other cells and to the developing organism as a whole. The Morgan school, however, apparently still believe that the cytoplasm is of secondary importance , 1 and that it is controlled by the genes . 2

While the theory of the gene in its present formulation has little significance for the major problems of heredity and development, dealing as it does in terms of abstract and hypothetical units, whose function it is to explain differences but not the ontogenetic process as a whole, there remains of course the possibility of a physiological or physico-chemical reinterpretation of the demonstrated facts which will throw light on the real functions of the chromosomes in general metabolism and development. That the chromosomes exert a powerful influence in modifying the course of development is already clear, and for this advance we have to thank the geneticists; it should not be long before some adequate physiological theory is formulated as to the way in which this influence is exerted. There are in fact not wanting signs that the gene theory is already in course of modification and, it may be, transformation. Something is already known of the actual chemical processes underlying the Mendelian inheritance of pigments, and the theory of genic balance 3

1 T. H. Morgan, ‘Genetics and the Physiology of Development*, American Naturalist , lx, 1926, pp. 489-515.

2 Morgan, Bridges, and Sturtevant, ‘The Genetics of Drosopbila\ Bibliograpbia Genetica y ii, 1925, p. 84. See also Morgan, Experimental Embryology , 1927, pp. 7-9, 208-9, and 654-5.

3 ‘Bridges* work on genic balance shows that the effect of a gene depends not only upon environmental conditions but also, and particularly, upon the other genes with which it is associated in the hereditary constitution of the individual*, Crew, op. dt.,

P- 733 3727



appears to be a step in the direction of a more physiological interpretation of the gene hypothesis. To become really physiological, however, the theory must throw off all traces of the particulate conception of heredity; having already shaken off the idea of representative particles, it must go farther and get rid of hypothetical ‘particles’ altogether. We shall at a later stage (see Chapter XIV) take up this important question of the relation of the chromosomes to development, and attempt to fit in the genetic facts with an ‘organismal’ theory of development . 1

To sum up with regard to the gene theory in its original formulation — we have seen that the theory is markedly inferior to Weismann’s as an explanation of development, that, in fact, it is not really a theory of development at all. It gives a formal explanation, based upon the behaviour of the chromosomes and their presumed constituent elements, the genes, of certain rather superficial aspects of heredity, namely the laws of inheritance of such minor mutations as are not incompatible with fertile inter-crossing, but it has not as yet linked up its ultimate elements, the genes, in any understandable way with our present knowledge of the physiology of the cell and the organism. It can by its very nature offer no explanation of the spatial and temporal harmony of development, nor of recapitulation, and it ignores completely the historical aspect of development. There is observable indeed in the writings of the Morgan school a distinct tendency to deny the validity of many of the concepts regarding adaptation and evolution which are commonly accepted by biologists — presumably because such concepts cannot find a place in the genetic scheme . 2

The methodological assumptions of the theory of the gene are very similar to those of Weismann’s theory. There is the

1 Our criticism of the gene theory, which may appear unnecessarily severe, is directed solely against its methodological assumptions. The value of the facts established by genetical research is not denied; the reader is referred to Chapter XIV (pp. 279-87) for an appreciation of the important part played by the chromosomes in development and heredity.

2 On the relation of the gene theory to the problems of evolution and adaptation, see J. T. Cunningham, Modern Biology, London, 1928.


same deep-rooted conviction that heredity must be explained as due to the transmission of an unchanging substance , the germ-plasm, in accordance with the materialistic principles which are accepted as the only basis for a scientific biology . 1 Like Weismann’s, the theory does not reach the stage of being stated in physico-chemical terms, but remains non-physiological and to a large extent abstract, though it is probably developing towards a more physiological formulation. Like Weismann’s again, it postulates independent material determinant units, though the simple relation between determinant and determinate is gone. These units, the genes, are arrived at by a process of reification of differences, and are in fact quite as hypothetical as Weismann’s determinants. Though the genes remain rather bare of qualities, as befits their function as symbols, it is assumed that they do in some way ‘determine’ the appearance of the characters of the organism. We shall consider this assumption more fully at a later stage.

1 One of the leading exponents of the gene theory writes: * . . . the biological investigator, in wielding his formulae, should not remain content until the abstract “tendencies” or “concepts” he arrives at can be translated analytically into terms of the arrangements and methods of movements of concrete particles’, H. J. Muller, American Naturalist , lxi, 1927, p. 416.